May 14, 2023
by Stephen Stofka
Asylum and AI
This week’s letter is about immigration from a historical perspective. This past Thursday marked the end of Title 42, the Covid-era policy that allowed border officials to quickly deport many immigrants who crossed into the U.S. at places other than official checkpoints. This generation of Americans is unlikely to come to a final resolution of the immigration issue that has plagued our politics for 150 years. The debate over asylum is more than 80 years old, first sparked by an incident just before the outbreak of World War 2. Can recent advances in machine intelligence help us resolve the bottlenecks created by our own Congress?
Most of us are familiar with Ellis Island, the immigrant processing center in New York Harbor. According to The Statue of Liberty – Ellis Island Foundation (2022), officials at the center processed 12 million immigrants during its 62 year history. The center was mostly active for just thirty years during that time. Most of those immigrants came in the years 1892 – 1924, when Congress passed an immigration law that limited admittance to people from mostly northern European states with family members already in the U.S. Across the river from Ellis Island was an earlier era in U.S. immigration – Castle Gardens at Battery Park in lower Manhattan. From 1855 – 1890, that center processed eight million people, mostly from those same northern European countries.
Like the immigrants appearing at our southern border today, most of those early immigrants came here for better economic opportunities. In countries across northern Europe, promoters of land offered farmland for sale in Nebraska and other Midwest states, recounted by Richard White (2017) in his The Republic For Which It Stands, a thorough history of the Reconstruction and Gilded Age periods of the 19th century. The harvests of wheat in the Midwest states helped drive down prices for wheat in Europe. Lower prices made farming less profitable in those countries and helped drive immigrants to the U.S.
Pundits like cartoonist Thomas Nast ridiculed the ethics employed by the U.S. government. It had pushed the Indians off their homeland, then sold that land to railroads at gift prices. Once the track had been built, the railroads marketed their surplus land to their future customers, immigrant farmers who would rely on the railroads to get their crops to market. Despite the American myth of the Midwest farmer, most immigrants became wage workers. When gold was discovered on Indian territory in South Dakota’s Black Hills, many European immigrants rushed toward the promise of riches and ignored the property rights of the Indians.
In California, voters had rejected ratification of the 14th Amendment on a slippery slope premise. If Negro men were given suffrage, American Indians, immigrants from China and southern Europe would soon be granted the right to vote and the country would be overrun with the mongrel races. Advocates for immigration reform attracted voter support based on these longstanding prejudices.
Following World War 1, two incidents prepared fertile ground for a coalition of immigration reformers to help pass restrictive immigration legislation. According to the U.S. Citizenship and Immigration Services (2023), the Quota Acts of 1921 and 1924 mostly excluded immigrants who were not from northern European countries. The first was the “Spanish” flu which originated in Kansas, not Spain. News of the disease’s spread was first published in Spain because other countries, including the U.S., suppressed publication. The belief that the disease had originated in southern Europe justified the prejudice against southern Europeans as being unclean (White, 2017).
The second incident was the severe 18 month depression of 1920-21. According to the Social Security Administration (n.d.), unemployment in various industries ranged from 14% in transportation to 27% in construction and 38% in mining. Immigrants competed with wage workers particularly in lower skilled jobs. Restricting immigration reduced those economic pressures.
The U.S. did not have an asylum policy until after World War 2. The precipitating incident came in 1939, when a ship loaded with almost a thousand Jewish refugees left from Germany for Havana, Cuba. Many passengers planned to wait in Cuba while their U.S. visa applications were approved, according to the United States Holocaust Museum (n.d.). However, Cuba backed out of an agreement to receive them. Although the ship sailed close to the Florida shore, U.S. officials did not allow the passengers to disembark and they returned to Germany. Canada and some other countries accepted some refugees but a third were detained and died in the Nazi concentration camps.
The incident tarnished America’s image. After the end of the war with Japan in August 1945, President Truman issued a directive that admitted some displaced persons to the U.S. (USCIS, 2023). It took Congress two more years to formalize an asylum arrangement with the passage of the Displaced Persons Act in 1948. Reflecting long held prejudices among Americans, Congress kept its quota system in place. In 1965, the Congress passed amendments to the 1952 Immigration and Nationality Act that finally abolished the quota system for refugees.
Today, the queue for refugee applications is several years long and many migrants claiming asylum wait in the U.S. until their case is resolved by an immigration judge. According to the USCIS (2023), two Congressional amendments in 1990 and 2004 reduced the burden of proof that migrant applicants must show to substantiate their claims of being political refugees. Those within and without the system admit that it is broken but an evenly divided Congress has not been able to resolve differences.
The development of ChatGPT has sparked a great deal of public interest in the capabilities of interactive Artificial Intelligence (AI) machines. Authorities at the border and in the courts are overwhelmed with migrants claiming asylum status. Most claims will be denied but the applicants get to work and stay in the U.S. while they wait. Might it be possible to use AI machines to process these claims? Some may object to the idea of machines controlling the destiny of vulnerable migrants. If machine intelligence is not adequate to safely navigate a car down a highway, can we trust them to make complex decisions regarding human safety and respect? For successful applicants, a quick decision would give migrants certainty and enable them to access job opportunities and government services that might otherwise not be available under the current system.
The machines could discover as much information as is possible in other countries to assess a migrant’s claims of political or criminal persecution. The machines could sort through the volumes of legal precedent that bog down our human decision-making. They could cite the relevant information and precedent that supported or did not support a claim of asylum. Applicants who were denied by an AI machine could appeal their claim but outside the country. It is not a perfect system but one that might be acceptable to advocacy groups on both sides of the issue.
////////////////
Keywords: immigration, asylum, AI, ChatGPT
Social Security Administration. (n.d.). Estimates of Unemployment in the United States. Social Security History. https://www.ssa.gov/history/reports/ces/cesbookc3.html
The Statue of Liberty – Ellis Island Foundation. (2022, November 1). Ellis Island. Statue of Liberty & Ellis Island. https://www.statueofliberty.org/ellis-island/overview-history/
U.S. Citizenship and Immigration Services. (2023, February 7). Refugee timeline. USCIS. https://www.uscis.gov/about-us/our-history/history-office-and-library/featured-stories-from-the-uscis-history-office-and-library/refugee-timeline
United States Holocaust Memorial Museum. (n.d.). Voyage of the St. Louis. United States holocaust memorial museum. https://encyclopedia.ushmm.org/content/en/article/voyage-of-the-st-louis
White, R. (2017). The Republic for which it stands: The United States during reconstruction and the gilded age, 1865-1896. Oxford University Press.
Innocent Investor 